Region-based Mixture of Gaussians modelling for foreground detection in dynamic scenes

作者：

Highlights：

• A generalised framework for region based mixture modelling is proposed.

• Batch and online update equations are derived using expectation maximisation theory.

• Region-based Mixture of Gaussians (RMoG) algorithm for dynamic background subtraction

• Experiments show state-of-the-art results in subtracting dynamic backgrounds.

摘要

Highlights•A generalised framework for region based mixture modelling is proposed.•Batch and online update equations are derived using expectation maximisation theory.•Region-based Mixture of Gaussians (RMoG) algorithm for dynamic background subtraction•Experiments show state-of-the-art results in subtracting dynamic backgrounds.

论文关键词：Region based modelling,Moving object detection,Mixture of Gaussians,Dynamic background subtraction,Expectation maximisation

论文评审过程：Received 15 December 2013, Revised 21 February 2015, Accepted 12 April 2015, Available online 24 April 2015, Version of Record 16 July 2015.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.04.016